The strong urge ofmanufacturing companies to become more flexible in their products and theimportance of assembly sequence planning place a high importance on characterizingproduct flexibility in an assembly system.
With little knowledge on this researchaspect, it is expected that it will contribute immensely to both theory andpractice by identifying the need for assembly sequence planning for flexibleproduct. This research will focus on assembly systems with semi-automatic assemblylines.Through assemblysequence planning, a feasible and optimal assembly sequence can be obtainedthrough which the parts can be assembled into the product successfully andefficiently with less assembly time or assembly cost. It has been seen thatthere are several works on assembly sequence planning with some limitations.Hence, this research study is proposing to utilize the generic algorithm basedapproach for assembly sequence planning for flexible product in which changestime of the assembly tools, assembly directions and assembly types will be usedin the fitness function to evaluate the assembly cost. Moreover, the influenceof tolerance and clearance on the product will be considered and non-dominatedsolutions will be found. More constraints will be considered to improve thestability in the assembly process which will help widen the capacity ofassembly sequence planning.
Case studies will be given so as to verify theproposed assembly sequence planning approach for flexible products. The work of Dini et al.(1999) proposed a method using genetic algorithms to generate and evaluate theassembly sequence, and adopted a fitness function considering simultaneouslythe geometric constraints and some assembly process, including the minimizationof gripper changes and object orientations, and the possibility of groupingsimilar assembly operations.
Hong and Cho (1999) proposed a GA-based approachto generate the assembly sequence for robotic assembly, and the fitnessfunction is constructed based on the assembly costs that are reflected by thedegree of motion instability, and assembly direction changes are assigned withdifferent weights. Lazzerini and Marcelloni (2000) used GA to generate andassess the assembly plans. The fitness function is constructed throughassigning different weights to three criteria: number of orientation changes,number of the gripper replacements, and grouping of similar assembly operations;and the different assembly planning results are derived through adjusting theweights in the fitness function in the experiments. Chen and Liu (2001)proposed an adaptive genetic algorithm (AGA) to find global optimal ornear-global-optimal assembly sequences. An original ordering generic algorithmto plan assembly sequence was proposed by Lit et al. (2001). Some success hasbeen achieved in the GA-based assembly planning works by Smith and Smith (2002)who proposed an enhanced genetic algorithm based on the traditional geneticalgorithm. For the research works on assembly sequence planning, Lu et al.
(2006) and Guan et al. (2002) proposed the assembly planning approaches withgenetic algorithm (GA), where the assembly sequences are regarded as chromosomes,and the solutions are evolved through crossover and mutation operation. Lv andLu (2010) and Wang and Liu (2010) proposed the particle swarm optimizationapproach to assembly sequence planning, and this approach is easier toimplement with the fewer computation procedures and fewer parameters. Lu andLiu (2012) proposed a disassembly sequence planning approach with an advanced immunealgorithm, by which the optimal or near-optimal assembly sequence can be derivedby converting the generated disassembly sequences. Lu et al. (2008) and Wang etal. (2005) proposed the ant colony optimization approach to disassembly planningor assembly planning.
The disassembly sequence or assembly sequence can be builtstep by step with the mechanism of the ant colony optimization, and the optimalor near-optimal sequences can be easily found.Assembly sequenceplanning is one of the best-known productions scheduling problems and proved tobe a strongly NP-hard problem. It has a focus of determining the order ofprocessing jobs in the assembly line, to save the assembly cost or shorten theassembly time. Recently, some artificial intelligence-based technologies havebeen utilized in the assembly sequence planning.
Knowledge-based approach andGeneric algorithm-based approach are the two areas in which artificialintelligence based approach can be divided. Although the mechanism ofknowledge-based approach can find feasible assembly sequence, however; whenassembly has many parts and components, and many alternative assembly sequencesexist, it is difficult to find optimal assembly sequence without an optimal searchalgorithm. Generic algorithm-based approach for assembly planning has receivedgreater research interest because both the optimal and near optimal solutioncan be found with high computing efficiency being achieved. Hence, the genericalgorithm is a promising approach to be utilized for this study. Assembly planning is animportant step during product development. Flexible product development is theability to make changes to the product being developed or how it is beingdeveloped without being too disruptive.
The main objective of assembly planningis to find a feasible assembly sequence with the minimum assembly cost andassembly time. To improve profit margin,effective assembly planning is important so as to significantly reduce theproduct development cost.